ChatGPT Just Removed Its Free Tier Limits — And Added a New “Thinking” Mode for Everyone
Today’s AI news circles around one clear theme: access. Who gets to use powerful AI tools, how much, and at what cost? Two of today’s top stories come from OpenAI expanding what free users can do, while a third shows an AI company trying to clean up a mess that open access helped create.
ChatGPT’s Free Version No Longer Has a Chat Limit
Starting next week, you’ll be able to open ChatGPT without paying a cent and talk to it as long as you want. OpenAI is removing the cap on text conversations for free and basic-tier users, according to The Verge. Until now, free users would hit a wall after a certain number of chats and either have to wait or upgrade to a paid plan.
Think of it like a coffee shop that used to give free Wi-Fi for only 30 minutes. Now the timer is gone. You can stay as long as you like, at least for text-based conversations. Voice features, image generation, and other premium tools still live behind a paywall.
For everyday people, this matters most if you’ve been using ChatGPT casually and kept bumping into that “you’ve reached your limit” message at the worst possible moment — mid-project, mid-draft, mid-thought. That friction is going away.
Why this matters: Removing limits lowers the barrier for millions of people who couldn’t justify paying for a subscription. More people using the tool means more people forming habits around it.
“Free users will be able to chat with ChatGPT as much as they want without hitting usage limits.”
Free Users Are Also Getting a “Think” Button for Harder Questions
Alongside the unlimited chats, TechCrunch reports that OpenAI is also rolling out a new “think” button for free and Go-tier users. This is a feature that tells ChatGPT to slow down and reason more carefully before responding, rather than firing off a quick answer.
AI models like ChatGPT normally generate responses very fast, essentially predicting one word at a time in rapid succession. The “think” mode asks the model to work through a problem step by step before it answers — similar to telling a friend to take a breath before they blurt something out. The result is usually more accurate answers to complex questions like multi-step math, legal scenarios, or detailed research tasks.
Previously, this kind of deeper reasoning was reserved for paid subscribers. Giving it to free users is a meaningful upgrade. If you’ve ever gotten a confident but wrong answer from ChatGPT on something complicated, this feature is designed to reduce exactly that problem.
Why this matters: Better reasoning on hard questions, available for free, makes ChatGPT more trustworthy for the kinds of tasks where mistakes actually cost you something.
“Removed limits on text conversations for free ChatGPT users.”
Suno Is Watermarking AI Music to Fight Spam
Suno, one of the most popular AI music generation tools, is quietly dealing with a side effect of its own success. The company is now adding invisible watermarks — hidden digital tags that identify a song as AI-generated — to every track it produces, and it’s tightening its download policies, according to The Verge. The goal is to reduce a flood of low-quality AI music being dumped across streaming platforms and the web.
A watermark in this context isn’t a visible logo. It’s a signal embedded in the audio itself that identifies the song’s origin, much like a serial number stamped inside a piece of jewelry. Platforms and detection tools can read it even when listeners can’t hear it. Changing the download rules makes it harder for people to grab songs in bulk and upload them elsewhere automatically.
For music fans and creators, this is a sign that the AI music industry is starting to reckon with its own pollution problem. Unlimited AI-generated content, produced cheaply and at scale, has been clogging platforms with filler tracks that push human artists further down recommendation algorithms.
Why this matters: Watermarking is one of the first concrete steps any AI music company has taken to be accountable for what its tool produces. It won’t solve everything, but it sets a precedent.
“Adding watermark to songs and changing download rules to combat spam.”
Also Happening in AI
OpenAI is reportedly developing an AI smart speaker priced between $300 and $400, according to TechCrunch, putting it in direct competition with Amazon’s Echo and Google’s Nest devices. On the software side, Ollama released version 0.32.6, which improves performance for Qwen3 models specifically on Apple computers — good news for anyone running local AI models on a Mac. OpenAI also quietly pushed out a minor update to its Python developer library, fixing a build issue that was tripping up developers. Elsewhere, Ars Technica published a sobering look at why AI moderation tools alone can’t protect social media communities from AI-generated content — a reminder that automated problems don’t always have automated solutions. Finally, a startup called Naïve raised $28.5 million to automate the administrative work of starting and running a business, from paperwork to operations.
What to Watch
OpenAI is making a serious push to grow its free user base right as competition from Google, Anthropic, and open-source tools is intensifying. Watch whether this access expansion translates into more paid conversions down the line — that’s almost certainly the business logic behind it. The Suno watermarking story is also worth following closely: if regulators or streaming platforms start requiring watermarks on all AI-generated audio, this quiet policy change becomes a much bigger industry standard.